Are Existing Monocular Computer Vision-Based 3D Motion Capture Approaches Ready for Deployment? A Methodological Study on the Example of Alpine Skiing.

Ostrek Ostrek,Spörri Spörri,Fua Fua,Rhodin Rhodin,Müller Müller

doi:10.3390/s19194323

Abstract

In this study, we compared a monocular computer vision (MCV)-based approach with the golden standard for collecting kinematic data on ski tracks (i.e., video-based stereophotogrammetry) and assessed its deployment readiness for answering applied research questions in the context of alpine skiing. The investigated MCV-based approach predicted the three-dimensional human pose and ski orientation based on the image data from a single camera. The data set used for training and testing the underlying deep nets originated from a field experiment with six competitive alpine skiers. The normalized mean per joint position error of the MVC-based approach was found to be 0.08 ± 0.01 m. Knee flexion showed an accuracy and precision (in parenthesis) of 0.4 ± 7.1° (7.2 ± 1.5°) for the outside leg, and −0.2 ± 5.0° (6.7 ± 1.1°) for the inside leg. For hip flexion, the corresponding values were −0.4 ± 6.1° (4.4° ± 1.5°) and −0.7 ± 4.7° (3.7 ± 1.0°), respectively. The accuracy and precision of skiing-related metrics were revealed to be 0.03 ± 0.01 m (0.01 ± 0.00 m) for relative center of mass position, −0.1 ± 3.8° (3.4 ± 0.9) for lean angle, 0.01 ± 0.03 m (0.02 ± 0.01 m) for center of mass to outside ankle distance, 0.01 ± 0.05 m (0.03 ± 0.01 m) for fore/aft position, and 0.00 ± 0.01 m2 (0.01 ± 0.00 m2) for drag area. Such magnitudes can be considered acceptable for detecting relevant differences in the context of alpine skiing.

Highlights

In the last decade, technological evolutions have significantly impacted various fields of science and daily life
One potential solution to overcome the limitation of requiring complex camera calibration procedures and extensive manual annotation efforts when using video-based stereophotogrammetry for in-field 3D human pose estimation might be found in a deep learning approach, as was proposed most recently in Rhodin et al [21]
We introduced the vector–based formulation is easier to compute from joint location estimates

Summary

Introduction

Technological evolutions have significantly impacted various fields of science and daily life This is true for the area of sports, where recent advances in measurement technology have enabled the extraction of a variety of novel performance-, load-, and health-related metrics. One potential solution to overcome the limitation of requiring complex camera calibration procedures and extensive manual annotation efforts when using video-based stereophotogrammetry for in-field 3D human pose estimation might be found in a deep learning approach, as was proposed most recently in Rhodin et al [21] This becomes feasible when there is a sufficiently large and representative manually annotated dataset to properly train the underlying algorithms. The basic idea behind this is to use existing video-based stereophotogrammetric datasets (i.e., manually annotated and calibrated multiple cameras views from different perspectives) for the training of a deep network

Objectives

Methods

Discussion

Conclusion